Study of Scheduling for Processing Real-Time Communication Signals on Heterogeneous Clusters

Abstract

Exploiting clusters to develop software radio systems is a much more promising research direction. When vast much feeble communication signal data collected from highspeed network enter a cluster, they will be partitioned and allocated to nodes for processing so as to realize the processing of high-gain and low-delay. This paper proposes a general real-time scheduler model for communication signals processing on clusters, a self-adaptive task partition method and a dynamic synthetical heuristic DSSA, which considers both deadline and makespan while scheduling. By tuning correlative parameters, the DSSA can achieve higher guarantee ratio. On the promise of satisfying timing constraints, the system also has higher throughput and better load balancing. We make comparisons between DSSA and DASAP, DALAP algorithms and the simulation results show that DSSA surpasses DASAP and DALAP.